NVIDIA Air offers the unique ability to simulate anything from a small network to an entire data center. Before you start configuration, routing, or management, consider the topology first.
A network topology is the layout or structure of how devices connect and communicate within a network. It describes both the physical arrangement and the logical flow of data.
Questions to consider include:
- Which operating system are you going with?
- How much storage and RAM should you allocate?
- What will you name your servers, and how will they connect?
NVIDIA Air provides several options for creating a network topology. This blog post discusses the advantages of using an external file that you can create, edit, and upload into NVIDIA Air. This method offers high scalability and reusability, opening the door for further automation with NVIDIA Air.
See the following video for step-by-step instructions on importing an external file into NVIDIA Air.
Simulating the massive scale of data centers
Large networks require tens or hundreds of servers. Each one must be named, allocated resources, and connected to your network as well as other servers. The drag-and-drop editor built into NVIDIA Air is great for smaller-scale simulations and testing. It works well for creating a visual reference that represents the abstract concept of building topologies, which are often difficult to visualize. However, it isn’t scalable enough for large use cases, such as the data center.
Using external files to represent a topology solves this use case. Instead of undergoing the tedious process of dropping a new node into the drag-and-drop editor and assigning its resources, even with the ability to copy nodes, you can quickly create many nodes in an easily modifiable file. This process is less cumbersome for large networks and reduces manual errors. The same applies to port connections, adding a NetQ server, and information about the out-of-band management network. It’s much easier to work with large-scale simulations using an external file.
Share, break, and rebuild
With external files, you can save your topology to reuse later or share it with a colleague to review and experiment. If you break your simulation later, you can easily rebuild it and start over using the same file without having to recreate the topology manually.
Along with collaboration, you can take advantage of a version control system, such as Git, for change tracking and reverting mistakes.
Almost every demo in the NVIDIA Air demo marketplace provides a GitLab code repository where you can find the topology file used to build the demo. These provide great examples for you to reference. You can review these for your custom topology or to learn more about how the demos are created. Click on the GitLab logo icon 
to view its code repository.
Automation
External topology files are a critical component of automation within NVIDIA Air. They’re your launch point for building new simulations and can be used in CI/CD pipelines for automation. With the help of the NVIDIA Air API or SDK, you can bring up new simulations using the topology file without accessing air.nvidia.com in your browser.
After building the topology, you can use scripting and automation tools such as Ansible, Python scripts, and ZTP to configure your simulation.
Get started
Learn how to create external topology files in the NVIDIA Air user guide. You can use any text editor you prefer, such as Notepad or VS Code.
When you’re ready, upload your file to Air using these upload instructions. Once uploaded, NVIDIA Air will build your topology, and you can make any further modifications using the drag-and-drop editor.
For an overview of NVIDIA Air and a step-by-step introductory video, see the An Introduction to NVIDIA Air blog post. You can take the accessibility of your simulations a step further by integrating services such as SSH and HTTP/S. Refer to our Connect Simulations with the Real World Using NVIDIA Air Services for more information.
